Dengue virus infections are an emerging mosquito-borne disease. A continuing global pandemic of dengue illness causes approximately 50-100 million infections each year with 3.5 billion people at risk for infection worldwide. Dengue is often a nonspecific febrile illness (dengue fever, DF);however, patients with dengue can manifest a more severe illness with bleeding tendency, thrombocytopenia, and plasma leakage (dengue hemorrhagic fever, DHF). Dengue illness has a substantial economic impact in tropical and sub-tropical areas and is associated with higher costs and longer disease duration than other febrile illnesses (OFI). There is an ongoing need to prevent and control the consequences of dengue infection- early monitoring of symptoms and signs may help guide treatment to limit the severity of dengue illness and may improve the utilization of limited hospital resources. There are unanswered questions about the dynamics of DF and DHF early in the onset of illness, as the widely used WHO classification system was not designed as a tool for prospective clinical management of suspected dengue cases. Also, patients with OFI may present with symptoms similar to dengue early in the illness. Our long-term objective is to establish correlative models and classification trees that can be used for distinguishing children with dengue from children with other febrile illness (OFI) and distinguishing mild dengue illness from severe dengue illness at an early (clinically relevant) stage of the disease process. The proposed research will use 12 years of systematically collected data from Thailand among children who presented with undifferentiated fever within the first 72 hours of fever onset. The Specific Aims of the proposed research are to: 1) describe and compare the temporal dynamics and patterns of clinical laboratory variables in Thai children with DF, DHF, and OFI, 2) establish and validate logistic regression models that can distinguish between DF, DHF, OFI, and severe dengue infections that warrant early hospitalization, and 3) establish and validate classification trees using data at the day of presentation to distinguish patients with DF vs. DHF, any dengue vs. OFI, and severe dengue vs. non-severe. An improved understanding of clinical laboratory variables associated with dengue and early classification of patients with suspected dengue may help to prevent the consequences of severe dengue infection and improve the management of dengue infections globally. PUBLIC HEALTH RELEVANCE: Dengue virus infections are an emerging global health threat and have substantial impact in tropical and subtropical resource-poor areas. There is limited understanding of dengue infections and the dynamics of the illness over time. This research proposes to gain better understanding of the disease process and establish and validate diagnostic tools that will be clinically usefully early in the onset of illness.